1138
Anal. Chem. 1985, 57, 1138-1144
(14) Hass, J. R.; Friesen, M. D.; Hoffman, M. K. Ofg. Mass Spectrom. 1979, 14, 9. (15) Mahle, N. H.; Shadoff, L. A. Blomed. Mass Spectrom. 1982, 9 , 45. (le) Hunt, D. F.; Harvey, T. M.; Russell, J. w. J . them. sot., Chem. Commun. 1975, 151.
(17) Lamparskl, L. L.; Nestrlck, T. J. Chemosphere 1981, 10, 3 . RECEIVEDfor review September 19,1984. Accepted December
17, 1984.
Assessment of Adsorption/Solvent Extraction with Polyurethane Foam and Adsorption/Thermal Desorption with Tenax-GC for the Collection and Analysis of Ambient Organic Vapors Mary P. Ligocki and James F. Paekow* Department of Chemical, Biological, and Environmental Sciences, Oregon Graduate Center, 19600 N. W. Von Neumann Dr., Beaverton, Oregon 97006
Two methods for the collection of ambient organic vapors at the ng/m3 to pg/m3 level were utllired In field sampling at a residential site in Portland, OR, during the winter and spring of 1984. The methods were adsorptlon/solvent extraction with polyurethane foam plugs (ASE/PUFP) and adsorption/ thermal desorption wlth Tenax-GC cartrldges (ATD/TenaxGC). ASE/PUFP was used with a single sample flow rate In a single channel of the sampler. ATD/Tenax-GC was used with two different sample flow rates In two separate channels. Each method was found to be well sulted to the analysis of compounds In a speclflc range of volatility. Some Intermediate-volatility compounds were determined with all three sampling channels. The coefficients of varlation for the three channels pooled over seven events were 9-38% for compounds In the range of volatility between acenaphthene and pyrene. The low sample volumes used wlth ATD/Tenax-GC for determinations at the ng/m3 level make It an attractive method for many applications.
Over the past 2 decades, increasing numbers of trace organic compounds have been identified in the atmosphere. These include the polycyclic aromatic hydrocarbons (PAHs) and their derivatives, pesticides, and polychlorinated biphenyls (PCBs), many Qf which are of concern due to their toxic and carcinogenic properties. In most cases, the distribution and fate of these compounds are only poorly understood. Since knowledge of the physical state of an atmospheric contaminant is vital for the understanding of its operative physical and chemical removal processes, and since the vapor pressures of many trace organics, including the PAHs, pesticides, and PCBs, fall in the range where substantial amounts of a given compound are expected in both the vapor and aerosol phases, a comprehensive organic air sampling system must include both aerosol and vapor measurement capabilities. A popular sampler configuration for the determination of atmospheric organic compounds includes a glass fiber filter followed by an adsorbent such as polyurethane foam (PUF) or Tenax-GC (1-3). P U F has the advantages of being convenient to handle and inexpensive but exhibits breakthrough of volatile compounds ( 4 ) . Although Tenax-GC exhibits less breakthrough, small cartridges have a high flow resistance and large adsorbent beds may be prohibitively expensive for many
Air Intake 1 0 2 mm g l a s s holder
[anodized aluminum1
steel ,Stainlass 1 1 4 ' tublng
H o l d e r f o r PUFPs (atalnless s t e e l )
Polyurethan
foam plugs (PUFPs)
a n d pump ( 6 0 0 mL1mln)
Figure 1. Air sampler designed for the collection of ambient particulate and vapor phase organic compounds.
applications. In addition, Tenax-GC may degrade during sampling in the presence of reactive gases such as O3 and NOz (5). Billings and Bidleman (4) compared PUF and Tenax-GC in high volume field sampling for PCBs. Both adsorbents retained the less volatile PCBs adequately, but some breakthrough was observed on P U F for the more volatile PCBs. Solvent extraction was used to recover the analytRs from both types of adsorbents in that study. In conjunction with a sampling program to determine the scavenging of atmospheric organics by rain (6),an air sampler was developed which has the capability to collect organic compounds ranging in volatility from trichloroethene to coronene, in both the vapor and aerosol phases. It utilizes adsorption/solvent extraction (ASE) with P U F for the determination of low volatility organics, with analysis by capillary gas chromatography/mass spectrometry (GC/MS). Adsorption/thermal desorption (ATD) with Tenax-GC is used with
0003-2700/85/0357-1138$01.50/00 1985 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 57, NO. 6, MAY 1985
analysis by capillary GC/MS for the determination of highto-intermediate volatility organics. The overall sensitivity of ATD allows small, relatively inexpensive Tenax-GC cartridges to be used with low sample flow rates, while still demonstrating a detection limit below 1ng/m3. While each of these methods is particularly suited for the collection of compounds of a specific range of volatilities, a number of compounds are collected with high efficiency by both methods. These compounds can be used t o provide an internal comparison of the ASE/PUF and ATD/Tenax-GC methods. Although both methods have been used individually by other researchers, this type of comparison has never been conducted. E X P E R I M E N T A L SECTION Materials. Glass fiber filters were from Gelman, Inc. (Ann Arbor, MI). Tenax-GC (35/60 mesh) was purchased from Alltech Associates (Deerfield, IL). Polyurethane foam of density 0.022 g/cm3 was purchased from Dayco Northwest (Portland, OR). All solvents were glass distilled (Burdick & Jackson, Muskegon, MI). Perdeuterated benzene, toluene, 1,2-dichlorobenzene,naphthalene, acenaphthene, fluorene, phenanthrene, anthracene, benzophenone, chrysene, and perylene were from KOR Isotopes (Cambridge, MA). Fluoranthene-dlo was from MSD Isotopes (Los Angeles, CA). Polychlorinated biphenyls were from Ultra Scientific (Hope, RI). Other standard materials were obtained from Chem Service (West Chester, PA). Apparatus. The air sampler (Figure 1)included primary and backup 102-mm glass fiber filters followed by three parallel vapor sampling channels. Over 99% of the flow continued directly through a polyurethane foam plug (PUFP) 7.6 cm long and 5.1 cm in diameter at a rate of 150 L/min. Flows of 40 mL/min and 600 mL/min were diverted through two air desorption cartridges, ADC-1 and -2, respectively. The ADC bodies were constructed of Pyrex glass, with 1.1cm o.d., bed length 8.0 cm, and bed volume 5.7 cm3, and packed with 0.79 g of 35/60 mesh Tenax-GC. The Tenax-GC was held in place with silanized glass wool plugs. The ends of the cartridges were of precision 0.25 in. (0.64 cm) 0.d. tubing, which fit into brass Swagelok fittings with Teflon ferrules. Each channel was equipped with backup adsorbent units. All adsorbents were shielded from light during sampling, transport, and storage. Prior to sampling, PUFPs were cleaned by Soxhlet extraction for 24 h in acetone and dried under a stream of prepurified nitrogen. PUFPs were stored and transported in clean screwcapped glass jars fitted with TFE Teflon cap liners. The Tenax-GC cartridges were extracted with 2 L of 60:40 acetone:: hexane, and then conditioned at 275 “C for 4 h under a stream of ultrapure helium. The cartridges were capped with precleaned brass Swagelok caps fitted with Teflon ferrules, and stored and transported in muffle-furnace-baked Pyrex culture tubes. The sampler was situated a t ground level in a residential section of southeast Portland, a t the Oregon Department of Environmental Quality (ODEQ) air monitoring station a t 5824 S.E. Lafayette St. A Gast 1022 oilless carbon vane pump (Gast Manufacturing Corp., Benton Harbor, MI), housed approximately 5 m downwind of the sampler, provided the overall flow rate. For this study, the pump was actuated by a rain sensor, allowing air sampling only during periods of rain. The total sampling time was recorded. The flow rate through each of the channels was measured with a laboratory-calibrated rotameter (Dwyer Instruments, Michigan City, IN). Analysis Methods. PUFP recovery studies were carried out using standards prepared in acetone containing all analytes of interest, as well as the internal standard compounds. For these studies, 400 ng per component was injected into a PUFP in 40 pL of acetone and then subjected to the extraction, separation, and concentration procedure described below for PUFPs used in sampling. After sampling, PUFPs were spiked with 50 pL of an internal standard in acetone of the following composition (ng/pL): acenaphthene-d,, (20); fluorene-d,, (20); benzophenone-d,, (20); phenanthrene-d,o (20); fluoranthene-dlo (20); chrysene-d,, (20); perylene-d,, (20);o,p’-DDE (1);and o,p’-DDD (1). They were Soxhlet extracted with 500 mL of 6040 acetone:hexane for 3 h
1139
Table I. Recoveries of Neutral Compounds from PUFP Extraction and Separation Procedurea int absolute rel, std recovery: % recovery,c %
compound
A. Polycyclic Aromatic Hydrocarbons and Related Compounds
acenaphthylene acenaphthene fluorene 9-fluorenone dibenzothiophene phenanthrene anthracene 1-methylphenanthrene 9-methylanthracene 9,lO-anthracenedione fluoranthene pyrene benz[a]anthracene chrysene benzo[b + klfluoranthene benzo[a]pyrene perylene
68f4 72f3 72f3 71f4 75f5 72f5 71f3 75428 74f8 72f9 78 f 10 82 f 11 81 & 10 84 f 12 96 f 17 90f 11 99 f 18
1 1 2
3 4 4 4 4 4 3 5 5 6 6 7 7 7
97 f 1 102 f 3 99 f 4 100 f 3 103 f 2 101 f 3 98 f 5 102 f 3 101 f 5 102 f 8 100 f 1 105 f 5
97 f 4 100 f 7 103 f 7 98 f 10 108 f 8
B. Pesticides and Polychlorinated Biphenyls
2,6-dichlorobiphenyl 3 4 a-HCH 4 hexachlorobenzene Y-HCH 4 2,5,2’-trichlorobiphenyl 4 2,5,4’-trichlorobiphenyl 4 4 heptachlor 2,5,2’,5’-tetrachlorobiphenyl 8 aldrin 8 2,5,3’,4’-tetrachlorobiphenyl 8 2,4,5,2’,5’-pentachlorobiphenyl 8 p,p’-DDE 8 dieldrin 8 3,4,3’,4’-tetrachlorobiphenyl 9 p,p’-DDD 9 2,4,5,2’,4’,5’-hexachlorobiphenyl 9 p,p’-DDT 9
71f3 71 f 10 74 f 10 72f8 75f6 73f8 6 8 f 11 77 f 12 78f7 78 f 14 82 f 17 85 f 19 81 f 13 81 f 15 77 f 20 87 f 20 80 f 22
102 f 4 97 f 10 100 f 6 99 f 7 103 f 12 100 f 3 93 f 7 103 f 8 106 f 15 104 f 3 109 f 2 110 f 4 109 f 6 102 f 6 97 f 2 110 f 3 99 f 4
68f4 72 f 10 75 f 14 78 f 16
97 f 10 103 f 12 96 f 7 100 f 9
79f8 89 f 10 102 f 22 274 f 50d
113 f 10 115 f 5 121 f 17 326 f 47d
C. Alkanes tetradecane hexadecane eicosane heneicosane
1
3 5 5 D. Phthalates
diethyl phthalate dibutyl phthalate butyl benzyl phthalate bis[2-ethylhexyl] phthalate
3
5 6 6
f l s values are based on three replicate samples. *Absolute recoveries are based on the external standard anthracene-d,,. cRelative recoveries are based on internal standards: (1) acenaphthene-d,,; (2) fluorene-dlo; (3) benzophenone-d,,; (4) phenanthrene-dlo; (5) fluoranthene-d,,; (6) chrysene-d,,; (7) perylene-d12; (8) o,p’-DDE; (9) o,p’-DDD. dNot included in average recovery calculation.
(at least 10 Soxhlet cycles). The extracts were concentrated to -20 mL in a Kuderna-Danish (K-D) apparatus and then further concentrated to 1 mL in a miniature K-D apparatus. The concentrated extracts were separated into acid and base/neutral fractions as follows. An extract was brought up to 10 mL in hexane, placed in a small separatory funnel, and extracted with two 15-mL portions of 0.01 N NaOH. The organic layer, containing the base/neutrals, was dried over 2 g of anhydrous Na&304. The aqueous layer was acidified to pH 2 with concentrated H,S04 and then extracted with two 15-mL portions of methylene chloride. The base/neutral fraction was again concentrated to 2 mL in the miniature K-D apparatus and then transferred to a 3-mL Mini-vial (Alltech Associates, Deerfield, IL)and the volume further reduced
1140
ANALYTICAL CHEMISTRY, VOL. 57, NO. 6, MAY 1985
Table 11. Percent Breakthrough" ( B )of Intermediate Volatility Organic Compounds on PUFPs during Sampling in 1984 in Portland. OR
2/ 14 53 m3 6 "C
2/12
compound hexadecane acenaphthylene acenaphthene dibenzofuran fluorene octadecane hexachlorobenzene a-HCH 9-fluorenone phenanthrene
int stdb
230 m3 8 "C
2 1 1 1 2
44 30 35 20 9 3 10
3 3 3 4 3
%B samDle date., vol., and mean temo 2/20 2/23 2/29 150 m3 200 m3 50 m3 5 "C 7 "C 9 "C
6 3
37 39 46 5
9 15
35
9.4
4.9
6.1 >3.5 3.6
1.3 5.3 3.3 4.4 21 22 22
3.9
12 17 NA 8.8 7.4 6.5 0.36 ND 0.33 0.09 ND 0.09
12
8.5 9.8
0.35 ND 0.32 0.05
concn, ng/m3 sample date 2/20 2/23 2/29 >27 3.4 2.1
>5.9 5.3 6.4 22 29
25 15 9.2 10 0.30
ND NDb 0.11 ND 0.07 6.7 6.4
2.8 5.7
ND ND 7.4 ND 10
1.1
1.2
0.6 1.4
ND 1.4
2.2 2.3 2.7
19
39
29
ND 26 12
ND 39 14
37 41 17 15 3.8 3.3 6.6 6.8 9.2
3.9
10
>15 1.7 0.82 >7.2 5.4 6.7 13 15 13
8.6 6.7 8.0 0.29 ND ND 0.07 ND ND 4.6 6.1 6.6 1.2 0.7
1.4 20 30 27 11
47 9.3 3.1 9.3
7.7 9.9 22 25 26 16 15
~
3ji6
4/11
pooled CV
>11 1.9
>14 0.9 NA >3.3 2.8 NA 14 16 NA 7.1 7.6 NA 0.44 ND ND 0.06 ND ND 4.2
139
NAB >4.2 5.9
NA 10
11
15
0.33 ND 0.43 0.07 ND NA 7.6 7.4 NA
16 NA 7.8 9.4 NA 0.33 ND ND 0.08 ND ND 4.0 15 6.4
1.6
1.8
1.4 NA 35 ND NA 18 NA 3.3 NA 5.4 NA
2.4 2.4 20 54 30
1.2 21 26 22
27
10
9.2
19
12
7.6 7.7 9,lO-anthracenedione 2.2 3.0 2.6 1.7 1.5 3.0 2.7 NA eicosane 2.9 2.1 3.6 3.9 3.2 2.3 3.9 3.9 fluoranthene 6.0 9.5 6.8 10 7.1 5.4 6.7 9.2 11 NA NA pyrene 5.4 6.5 10 8.4 5.8 8.5 4.1 6.7 4.8 NA 8.6 NA "NA = not available. bND = not detected at a significance significant level (95% confidence level). 8.3
16
17 13
36
6.9
7.6 1.2
26
1.0
9.9 1.5
NA 2.8 3.7 5.3 6.5 4.7 6.1
14 9
9 17
1142
ANALYTICAL CHEMISTRY, VOL. 57, NO. 6, MAY 1985
Table V. Levels of Benzaldehyde and Acetophenone Found on Field Blank, Backup Sample, and Primary Sample Tenax-GC ADC-2s in 1984 in Portland, OR" sample
benzaldehyde acetophenone date (ng/g Tenax-GC) (ng/g Tenax-GC)
laboratory blank
2.2
field blank field blank field blank field blank av field blank
2/14 2/23 3/16 4/11
94 60 52 66 68 f 18
backup backup backup backup backup backup av backup
2/12 2/14 2/20 2/23 2/29 4/11
290 160 120 93 39 81 130 f 88
primary primary primary primary primary primary primary av primary
2/12 2/14 2/20 2/23 2/29 3/16 4/11
2400 1000 1300 1100
730 1600 670 1300 f 600
4.1 430 160 95 120 200 f 150 180 290 79 120 47 130 140 f 86 870 610 520 500
130 1400 640 670 f 390
"The levels of NO, at the sampling site were measured by ODEQ and were "not available", 24, 32, 22, 33, 34, and 20 ppb for 2/12, 2/14, 2/20, 2/23, 2/29, 3/16, and 4/11, respectively. The summed concentrations for a total of 74 organic compounds were 10, 16, 18, 9.4, 33, 11, and 11 pg/m3 for the same dates, respectively.
to simulate the low amounts expected in the sample extracts, the P U F P recovery determinations were carried out so as to produce a final concentrated extract level of 2 ng/pL per component. The overall recoveries of members of several classes of organic compounds are shown in Table I. Excluding the phthalates, absolute overall recoveries of all compounds for the extraction/separation procedure were in the range 68-99%, and averaged 77%. Relative recoveries based on the internal standards were all in the range 93-110%. The standard deviations of the relative recoveries averaged 4% for the PAHs, indicating good analytical precision. The standard deviations of the relative recoveries of the other classes of compounds were larger, up to 12% for hexadecane and 15% for aldrin. I t is likely that this was due in part to the lack of use of closely related internal standards for these compounds. Phthalates are known to be troublesome compounds to determine at trace levels because of their ubiquitous presence in the laboratory environment. Table ID shows that only the bis(2-ethylhexyl) phthalate determination suffered from a severe contamination problem a t the 2 ng/pL level. Breakthrough of Analytes during Sampling. Seven sets of air concentration data were obtained during rain events in Portland in the winter and spring of 1984. Total sample volumes of 50-230 m3 were obtained over periods of 1-5 days. The average ambient temperature was 8 "C. In all concentration determinations, mean blank values were calculated from the mass amounts found on the blanks. Primary and backup sample amounts were considered nonzero only if they exceeded the mean blank amount at the 95% confidence level. Normalized blank levels for each event were then calculated by dividing the blank mass amounts by the corresponding sample volumes, and were then subtracted from the primary and backup levels. These normalized blank levels were generally less than 5% of the sample levels for the PUFPs and ADC-29, except in the case of the phthalates. Normalized
blank levels were higher for the ADC-1s due to the extremely low sample volumes collected. Breakthrough of an analyte on an adsorbent is a function of the ambient temperature, the sample volume, the adsorbent volume, the affinity of the analyte for the adsorbent, and the dimensions of the adsorbent trap. When backup adsorbent traps are utilized, the amount of material found on the backup trap provides an indication of the efficiency of the trapping process. By use of the breakthrough curves from Senum (8), and assuming 7.5 theoretical plates per P U F P (9), for an overall trapping efficiency of 99% on the two-plug system, the primary plug must retain a t least 75% of the influent material. Breakthrough (B, %) is defined here as [(amount of analyte on the backup cartridge) X 100%]/(sum of the amounts on the primary and backup cartridges). Thus a B value of